🔭 astrophysics

Detecting redshift space distortions with forthcoming 21cm intensity maps from MeerKLASS

This paper demonstrates through simulations that the MeerKLASS 21cm intensity mapping survey can achieve a significant $6.7\sigma$ detection of redshift space distortions and effectively constrain modified gravity parameters by employing a robust analysis pipeline that accounts for observational effects and leverages tomographic modeling to resolve key cosmological degeneracies.

Original authors: Daniel Tassie, Steven Cunnington, David Bacon, Zhaoting Chen, José Luis Bernal

Published 2026-10-12✓ Author reviewed ⓘ

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel Tassie, Steven Cunnington, David Bacon, Zhaoting Chen, José Luis Bernal

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The universe is not a static backdrop; it is a dynamic stage where matter clumps together under the relentless pull of gravity. Over billions of years, this process has woven a vast cosmic web of galaxies, clusters, and empty voids. To understand how this structure formed and grew, cosmologists look for subtle clues in the way matter is distributed. One of the most powerful clues lies in the motion of galaxies. While the universe is expanding, pulling everything apart, gravity also pulls matter together, causing galaxies to drift toward dense regions. This extra motion, known as a peculiar velocity, distorts our view of where galaxies actually are. When we map the universe using light, these motions create a specific pattern of stretching and squeezing along our line of sight. By measuring this distortion, scientists can determine how fast cosmic structures are growing, which in turn tests the fundamental laws of gravity that govern our existence.

A team of researchers has now demonstrated a promising new way to capture this signal using a powerful radio telescope in South Africa. Instead of looking at individual galaxies, which can be difficult to count at vast distances, they propose using a technique called intensity mapping. This method measures the total glow of neutral hydrogen gas, the most abundant element in the universe, across large patches of the sky. This gas emits a faint radio signal that shifts in frequency as the universe expands, allowing astronomers to map its distribution in three dimensions. The researchers focused on the MeerKAT telescope, specifically its Large Area Synoptic Survey, which is designed to scan the sky in a specific range of radio frequencies. Their work shows that despite the immense challenges of filtering out bright interference from our own galaxy and the telescope's own limitations, it is possible to isolate the subtle signature of cosmic growth. Through detailed computer simulations, they proved that they can recover the distorted signal with high statistical confidence, paving the way for future tests of gravity using this new observational tool.

The primary challenge in this endeavor is that the faint signal from the hydrogen gas is drowned out by much brighter radio emissions from our own galaxy and other sources. These foreground contaminants are thousands of times brighter than the cosmological signal the scientists want to measure. To solve this, the team simulated the process of removing these bright signals using a statistical method that identifies and subtracts the smooth, predictable patterns of the interference. However, this cleaning process is not perfect; it inevitably removes some of the desired cosmic signal along with the noise, effectively dimming the map. The researchers developed a mathematical correction, known as a transfer function, to reconstruct the lost signal. They tested this method on simulated maps covering a significant portion of the sky, equivalent to 1,115 square degrees, and found that the correction successfully recovered the distorted pattern caused by cosmic growth. The result was a clear detection of the signal at a level of 6.7 times the standard deviation, a threshold that indicates a very high level of certainty that the signal is real and not just random noise or an artifact of the telescope.

Beyond simply detecting the signal, the study explored what this method could achieve with the full power of the MeerKAT survey, which aims to map 10,000 square degrees of the sky. The researchers investigated how well they could measure the growth rate of the universe across different epochs, from when the universe was about half its current age to when it was nearly as old as it is today. They found that while the telescope's primary beam, which acts like a lens, naturally blurs the image and creates its own distortions, these effects could be distinguished from the true cosmic signal. By combining data from multiple redshift bins, or slices of cosmic time, and using a model that accounts for how the density of hydrogen gas changes over time, they could break down the uncertainties that usually make these measurements difficult. Their forecasts suggest that with the full survey, they could constrain the growth of structure with a precision that rivals other major cosmological experiments, offering a competitive test for theories of gravity.

The ultimate goal of these measurements is to see if gravity behaves exactly as Albert Einstein predicted or if it requires modification on the largest scales. The researchers modeled the growth of the universe using a parameter that describes how fast structures form. They found that if they assume the density of hydrogen gas and its clustering properties follow a smooth, predictable evolution, they can measure this growth rate with a precision of about 22.6 percent. This level of accuracy is sufficient to distinguish between standard gravity and some alternative theories that propose gravity behaves differently in the vast emptiness of space. However, the study also highlights a crucial dependency: the precision of these results relies heavily on having good prior knowledge of how much hydrogen gas exists in the universe and how it clusters. Without this external information, the measurements become much less certain. The work suggests that as our understanding of the hydrogen distribution improves through better simulations and observations, the MeerKAT telescope will become an even more powerful instrument for probing the fundamental nature of gravity.

This research represents a significant step forward in the field of radio astronomy and cosmology. It moves beyond the theoretical possibility of using intensity mapping to detect cosmic distortions and provides a concrete, validated pathway for doing so in real data. By successfully simulating the entire process—from the initial contamination by bright foregrounds to the final reconstruction of the cosmic signal—the team has shown that the signal of cosmic growth can be salvaged and measured with high confidence. The ability to detect this signal over the background noise of the telescope itself is a critical validation of the method. While the current results are based on simulations, they provide a robust blueprint for analyzing the actual data that MeerKAT is currently collecting and will collect in the coming years. The findings indicate that the next generation of radio surveys will not only map the distribution of matter in the universe but will also serve as a precise laboratory for testing the laws of physics that shape our cosmos.

Technical Summary: Detecting Redshift Space Distortions with MeerKLASS

Problem Statement
Redshift space distortions (RSD) offer a critical probe of the growth rate of cosmic structure, $f(z)$ , allowing for tests of General Relativity on large scales. However, detecting RSDs in 21cm intensity mapping (IM) surveys faces significant challenges. Unlike optical surveys, 21cm IM measures the total radio emission, which is dominated by foregrounds (galactic synchrotron, free-free emission, and point sources) that are orders of magnitude brighter than the cosmological signal. Standard "blind" foreground removal techniques, such as Principal Component Analysis (PCA), rely on the spectral smoothness of foregrounds but inevitably project out correlated modes along the line-of-sight (LoS), leading to cosmological signal loss. Furthermore, instrumental effects, specifically the frequency-dependent primary beam of the MeerKAT telescope, introduce artificial anisotropies and dampen small-scale transverse modes. These effects can mimic or suppress the RSD signal, creating degeneracies between cosmological parameters (such as the growth rate $f$ and the HI bias $b_{HI}$ ) and nuisance parameters. While transfer functions have been validated to correct signal loss for the monopole moment of the power spectrum, their efficacy in reconstructing higher-order multipoles—specifically the quadrupole, which carries the RSD information—remains unproven.

Methodology
The authors utilize simulated sky observations from the MeerKAT Large Area Synoptic Survey (MeerKLASS) covering the redshift range $0.4 < z < 1.45$ (UHF band).

Simulations: Using the meer21cm package, the authors generate 100 log-normal mock realizations of HI intensity maps, incorporating realistic foregrounds (based on PySM3), thermal noise, and the frequency-dependent primary beam of MeerKAT.
Foreground Cleaning: A "blind" PCA cleaning strategy is applied to remove foregrounds, with $N_{fg}=12$ modes removed to simulate aggressive cleaning levels required for current MeerKLASS analyses.
Signal Reconstruction: To correct for the signal loss induced by PCA, the authors employ a foreground transfer function, $T(k)$ . This is constructed by injecting mock HI signals into the observed data, re-cleaning, and calculating the ratio of the cross-power spectrum to the auto-power spectrum. This formalism is extended from the monopole to the quadrupole moment ( $P_2$ ).
Analysis Framework: The study employs a multipole expansion formalism to model the anisotropic HI power spectrum, $P(k, \mu)$ , including the Kaiser effect and Fingers of God (FoG) damping. The analysis focuses on the quadrupole to isolate RSD effects.
Forecasting: The authors perform Fisher forecasts for the full 10,000 deg $^2$ MeerKLASS survey. They explore parameter degeneracies between $f$ , $b_{HI}$ , and the cosmic HI density $\Omega_{HI}$ . To break these degeneracies, they adopt a tomographic approach across six redshift bins, modeling the amplitude parameters ( $b_{HI}$ and $\Omega_{HI}$ ) as redshift-dependent polynomials. They also test the growth rate index $\gamma$ as a probe of modified gravity.

Key Contributions

Validation of Quadrupole Reconstruction: The paper demonstrates, for the first time in 21cm IM, that a foreground transfer function can accurately reconstruct the quadrupole moment of the HI power spectrum after aggressive blind foreground cleaning.
Separation of RSD and Beam Anisotropies: The study quantifies the anisotropic signal induced by the primary beam and demonstrates that the RSD-induced quadrupole can be distinguished from this instrumental signal.
Tomographic Degeneracy Breaking: The authors show that while the HI auto-power spectrum suffers from strong degeneracies between $f$ , $b_{HI}$ , and $\Omega_{HI}$ , a tomographic analysis with shared polynomial amplitude parameters can partially alleviate these issues, provided external priors are applied.
Modified Gravity Probes: The work outlines a pathway to constrain the growth rate index $\gamma$ using MeerKLASS data, treating it as a simple probe for deviations from General Relativity.

Results

Detection Significance: Using 1,115 deg $^2$ of simulated MeerKLASS data ( $0.4 < z < 1.0$ ), the proposed pipeline achieves a 6.7 $\sigma$ detection of an RSD-induced quadrupole relative to a null model containing only anisotropic observational effects (beam and re-gridding artifacts). This confirms that the RSD signal survives aggressive foreground cleaning and can be recovered via the transfer function.
Forecast Constraints (Single Bin): For a single redshift bin at $z \approx 1.1$ , constraints on the RSD parameter $\beta = f/b_{HI}$ are limited by priors on $b_{HI}$ and $\Omega_{HI}$ . Without tight priors, the parameters are perfectly degenerate.
Forecast Constraints (Tomographic): By combining six redshift bins ( $0.4 < z < 1.45$ ) and modeling amplitudes as a quadratic polynomial, the constraints on $\beta$ improve, ranging from $\sigma(\beta)/\beta \approx 6.7\%$ at low redshift to $28.0\%$ at high redshift.
Modified Gravity Constraints: When parametrizing growth via the index $\gamma$ and imposing 10% Gaussian priors on the polynomial coefficients of $b_{HI}(z)$ and $\Omega_{HI}(z)$ , the combined tomographic analysis yields a constraint of $\sigma(\gamma)/\gamma = 22.6\%$ . This precision is comparable to current constraints from DESI peculiar velocity surveys.
Impact of Systematics: The inclusion of realistic scale cuts (to avoid beam-damped modes and wide-angle effects) degrades constraints, particularly at high redshifts where the beam size is larger. However, the tomographic approach mitigates this loss by sharing information across bins.

Significance and Claims
The authors claim that this work establishes a viable pathway for detecting RSDs in 21cm intensity mapping, a feat not yet achieved with MeerKLASS or similar experiments. The primary significance lies in validating that the quadrupole moment—the carrier of RSD information—can be reconstructed with high statistical significance despite the signal loss inherent in blind foreground cleaning.

The paper modestly concludes that while MeerKLASS holds the potential to place competitive constraints on modified gravity (specifically testing the DGP model or deviations in $\gamma$ ), this capability is heavily dependent on external knowledge of the HI bias and density. The authors emphasize that without informative priors on $b_{HI}(z)$ and $\Omega_{HI}(z)$ , the growth rate $f(z)$ remains effectively unconstrained due to amplitude degeneracies. They suggest that future improvements in hydrodynamical simulations and cross-correlations with optical galaxy surveys (e.g., DESI) are necessary to fully realize the potential of MeerKLASS and the future SKAO for precision cosmology and gravity tests.

More like this